Neurodynamic algorithms for constrained distributed convex optimization over fixed or switching topology with time-varying communication delay

Based on the distributed optimization operation mechanism of multi-agent systems, this paper considers how to solve the distributed convex optimization problem with linear inequality constraints when agents in the system have time-varying delays in information interaction. Neurodynamic algorithms which can counteract the negative effect of communication delay on the system are proposed and analyzed, respectively, over fixed or switched topology. In addition, sufficient conditions of linear matrix inequality form are given by the Lyapunov theory. Finally, the convergences and effectiveness of the proposed neurodynamic algorithms over fixed topology and switched topology are verified by numerical examples and a practical application.